一项新的研究论文探讨了大型语言模型（LLMs）中存在的文化偏见，发现与普遍认为的西方偏见相反，这些模型表现出对日本文化的明显偏好。该研究利用了一个名为CROQ（Culture-Related Open Questions）的新数据集来分析LLM的响应。研究人员观察到，当使用英语等高资源语言进行提示时，LLM会提供更多样化的输出，并且文化偏见似乎出现在监督微调阶段，而不是预训练阶段。 AI
影响 强调了大型语言模型中潜在的文化盲点，表明需要更多样化的训练数据和评估方法。
排序理由 学术论文，使用新颖的数据集分析大型语言模型的偏见。
AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →
LLMs have been showing limitations when it comes to cultural coverage and competence, and in some cases show regional biases such as amplifying Western and Anglocentric viewpoints. While there have been works analysing the cultural capabilities of LLMs, there has not been specifi…
LLMs have been showing limitations when it comes to cultural coverage and competence, and in some cases show regional biases such as amplifying Western and Anglocentric viewpoints. While there have been works analysing the cultural capabilities of LLMs, there has not been specifi…